The purpose of this project is to address functional endpoints of patients and their caregivers by examining the comparative effects of two psychoeducational interventions on various aspects of caregiver and patient well-being. The efficacy of skill training classes and information/support classes has not been well-documented in the long term care literature. In addition, this research will attempt to specify the psychological mechanisms responsible for observed outcomes by measuring self-efficacy expectations and relating these to outcome. It is expected that perceived self-efficacy will improve differentially across treatment conditions, and that self-efficacy perceptions will be strong predictors of positive caregiver change. Specifically, we expect that caregiver reported burden will decrease, and that ability to tolerate behavioral problems in the frail elder relative will increase, as a result of participation in either of the two active interventions (compared to "treatment as usual" which is here defined as participation in either inpatient respite, outpatient respite, or geriatric day care alone). We also expect that caregivers' quality of life will improve (i.e., depression will decrease, use of adapttive coping strategies will increase, and life satisfaction will increase). In addition, we will examine whether or not change in caregivers is related to change in functional status of the frail relative, and/or to decrease in the frequency of manifest behavioral problems. We will attempt to specify factors that seem strongly implicated in the decision to place an elder relative in a long term care institution, as well as variables that predict "dropout" vs. "completer" status during the intervention phase of this study. Methodologically, we have instituted experimental control regarding frail elder' functional capacities and the program from which they are recruited by blocking on both these variables, then randomly assigning caregivers to experimental conditions. Since previous intervention studies have generally lacked adequate control groups, the proposed design appears to represent a unique contribution to the field of long term care research.